Functional analysis is an experimental procedure used to identify environmental factors that are maintaining a behaviour of concern. In essence, it seeks to determine the function of a behaviour by directly manipulating antecedents in a highly controlled environment, to ensure if the hypothesized environmental variables are functionally related to the behaviour of concern. However, ethical and practical challenges are presented when the behaviour of concern is actively harmful to the individual and/or their surroundings. For instance, assessing self-injurious behaviour requires evoking said behaviour in a controlled environment, thereby running the risk of its reinforcement. Therefore, it is crucial to weigh the benefits of identifying the function of the behaviour against the risks of temporarily increasing the probability of its occurrence.

 

Najdowski et al. (2008) discussed possible solutions to mitigate these risks, such as a shortened assessment time and the use of protective equipment. However, these carry limitations, too. Brief assessments can yield inconsistent results in identifying variables maintaining the behaviour of concern while protective equipment has shown to alter behaviour, thereby affecting the validity of the results. 

 

To address these limitations, the authors conducted functional analyses of precursor behaviours, responses that reliably precede a behaviour of concern and can signal its imminent occurrence (like screaming or threatening before an act of aggression). Although these are topographically different to behaviours of concern, they are functionally equivalent if maintained by the same consequence, thus belonging to the same response class. The rationale of this approach is based on the idea that designing interventions to treat precursor behaviours belonging to a particular response class may mitigate the occurrence of behaviours of concern belonging to the same response class.

 

For their study, three individuals were referred for assessment and treatment of behaviours of concern. Brent, an 8 year old boy with a traumatic brain injury, was referred for aggression towards teachers and peers. Pete, a 5 year old boy with developmental delays, was referred for aggression toward teachers. Lastly, Tom, a 45 year old man with an intellectual disability, was referred for inappropriately grabbing others’ genitals and masturbation, at his day program.

 

Functional analyses were conducted 3 to 5 days a week, in naturalistic settings and the behaviours of concern were then operationally defined. Aggression was defined as hitting, kicking, biting, and scratching (further categorized as contact between the participant’s hands, legs, teeth, and nails and another individual’s body) and throwing objects. Masturbation was defined as the contact between the participant’s hand and genitals, either on or underneath their clothing. Inappropriate grabbing involved contact or attempted contact between the participant’s hand and genitals of another person. Through caregiver interviews and systematic observations, precursor behaviours were also identified. Brent was assessed to exhibit whining, crying, and threatening to hurt himself or others as precursor behaviours. Pete’s precursors were whining, crying, and statements about toys possessed by others. Lastly, Tom’s precursor behaviours were uttering sexual statements about a person’s genitalia.

 

A multielement design was used to analyze the precursor behaviours. Sessions lasted for 5 minutes or until precursor behaviours were apparent to prevent escalation. Each participant was placed under at least three conditions; attention, demand, and control play conditions. In addition, Pete and Tom were placed under tangible and alone conditions, respectively. Despite the presence of teachers and aides, only experimenters interacted with Brent. Pete’s sessions included a confederate partner due to his peer-directed aggression, and attempts to harm the confederate were blocked. Experimenters were present for all but the alone condition for Tom. Trained observers independently recorded data using handheld devices, and responses were calculated per minute. Interobserver agreement was assessed in a portion of sessions using 10-second interval-by-interval comparisons. Overall, agreement was consistently high across functional analysis, baseline, and treatment sessions for precursor behavior, severe problem behavior, and alternative responses.

 

Results from the functional analyses showed differentiated responding in one condition for all participants. For Brent and Tom, precursor behaviours were highest during the attention condition, suggesting an attention-maintained function, while Pete showed differentiation in the tangible condition, indicating his behaviour was maintained by access to tangibles.

 

Individualized Functional Communication Training (FCT) programmes that reinforced alternative responses, or mands, were implemented for each participant. Brent was instructed to raise a card if he needed to communicate while Tom was given a pager to signal available staff. Pete was instructed to ask for toys appropriately, every instance of which was reinforced by promptly handing him the toy. In addition, a thinning schedule of reinforcement was also implemented in Pete’s and Tom’s FCT interventions. If the experimenter present during Pete’s session had his name on a board, only then were his alternative mands reinforced with the toy. Similarly, Tom’s pages were only reinforced if each page was separated by 3 seconds, latency which was gradually increased to 10 minutes as the sessions progressed. Each session lasted at least 5 minutes, with Pete’s sessions increased to 10 minutes and Tom’s increased to 30 minutes to accommodate the reinforcement schedule. Treatment effects were assessed using an ABAB reversal design, to evaluate treatment effects.

 

Intervention results revealed a positive effect of the treatment on all participants. Baseline phases were characterized by low alternative mands and high precursor behaviour occurrences. However, during FCT phases, precursor behaviours reduced significantly while alternative mands were consistently high. Post-treatment follow-ups revealed no occurrences of behaviours of concern for Pete and Brent a week after the treatment. Tom’s behaviours of concern were also reported to be significantly reduced in a one-month follow-up post treatment.

 

Although the authors were unable to definitively explain the reduction in behaviours of concern, they hypothesized that reinforcing precursor behaviours during functional analysis and baseline of FCT may have caused its differential reinforcement, thereby resulting in its nonoccurrence. Regardless, their results were consistent with supporting literature; reinforcing a behaviour that occurs earlier in the hierarchy of a response class can suppress responses of behaviour further in the hierarchy. Despite the treatment effects, the authors highlighted certain limitations. Speculations about the precursor behaviours being possibly unrelated to the behaviours of concern due to the absence of relevant antecedents were made. In addition, the shortlisted precursors were accounted for through interviews with caregivers and teachers, thereby lacking a formal investigation in their relationship with the behaviours of concern. Probe sessions to assess said relationships were proposed as a strategy for future researchers. The authors characterized their study as a preliminary evaluation and encouraged future research to explore the effects of precursor-based procedures on behaviours of concern.

​

Citation:

​​

Najdowski, A.C., Wallace, M.D., Ellsworth, C.L., MacAleese, A.N. and Cleveland, J.M. (2008), Functional Analyses and Treatment of Precursor Behaviour. Journal of Applied Behavior Analysis, 41: 97-105. https://doi.org/10.1901/jaba.2008.41-97​